Gresham Computing
Operations departments are drowning in data, but automated number crunching can reduce risk, says Neil Vernon of Gresham Computing
How does Gresham approach data challenges in the financial services industry?
We wanted to introduce something new to the back and middle offices that could manage data integrity and control, and provide insight into the lifecycle of a transaction. About five years ago, we noticed that regulatory changes meant there were a lot of new controls required, particularly in the middle office.
Most of the controls were being managed on spreadsheets, and there was no proper enterprise software available to the middle and back office that would allow organisations to respond properly to the changing demands.
With a spreadsheet model, if something goes wrong in the organisation it can take days or weeks before anyone realises something is not quite right with the trade. The longer it takes to check a transaction’s integrity, the more costly it becomes to fix it, the more uncertainty there is, and the longer it takes. Where customers are involved, they can easily become exposed to the fact that not all of the internal systems are singing from the same song sheet, and that can affect customer confidence.
You can pay $10 to FedEx to carry a package from New York to London, and at every step you will be able to see where that package is. It’s generally correct—or, at least, gives the impression of being correct—and it generally arrives on time. However, if you make a $10 million transaction with a major organisation, that money effectively disappears into a black hole. It will eventually pop out on the other side, but there is no real guarantee that it will be correct.
Without this level of insight, organisations run the risk of misreporting what they have traded. For example, there have been occasions where data has been ‘summarised’ to the degree that banks couldn’t even tell a regulator whether they’ve been buying or selling commodities. That sounds unbelievable, but it’s because some of these transactions are so complicated, with so many attributes and so much complexity around the meaning of each attribute, that when you try to combine that data into a single view, most software fails. That means, of course, it all ends up in a spreadsheet, because spreadsheets can be a million columns wide. The problem is that some poor employee then has to go through all of that data, collating it into a report for the regulators, with strict time constraints. Obviously, there is a major risk that they will get it wrong—with this in mind, mistakes were bound to happen. We offer tools that allow organisations to apply the right controls, and to gain certainty around that data.
How does the Clareti Transaction Control product work?
The technology is typically wired up to one of the front-office systems so it receives transactions as soon as they appear, and also attached to middle- and back-office systems. It means all the systems involved in processing a transaction are aligned for a consistent view, creating a simple end-to-end comparison tool across a whole range of attributes.
If a transaction is described using 1,000 attributes, Clareti Transaction Control (CTC) generates a data model that can hold 1,000 attributes. If it only requires 50, it generates a model that can hold 50. There is no pre-defined scheme for clients to map on to. The technology then notes the type of transaction and the attributes it will require, and publishes an accurate report.
If the data is not consistent, CTC will raise a red flag. For example, if the front office disagrees with the middle office, it will prompt a manual investigation. That means the manual effort is on just 1 or 2 percent, on an exception basis, rather than 100 percent, of the transaction reporting process.
How easy is it to identify where an error has occurred?
At the very minimum, we can show which systems are agreeing and which are not, and this can be the starting point for root cause analysis. Then, at a meta level, we can show that, for example, 20 percent of issues in transaction reporting are happening between two particular systems.
We can ‘guide the surgeon’s knife’ in terms of the processes and the systems that are failing in their integrity. If a significant proportion of errors are coming from one place then that is where the institution should be spending its resources.
Are banks quick enough to embrace innovation?
What is actually happening is banks are innovating, and that’s great. New and interesting financial products are good for all of us. But the core platforms often don’t understand these new products, so institutions end up bolting on ‘helper’ applications that are not thorough enough and that are not allocated an appropriate budget.
The problem with new innovation is that it’s hard to know whether anything is going to work or not, so banks are not going to spend millions on a new type of trade if they’re only going to complete six of them. When they’re trading 6,000 they have to start paying attention, and when they’re trading six million they really need to get it right. Throughout that innovation curve institutions are carrying an awful lot of risk. That’s when systems go wrong.
The more mainstream those innovative products become, the more volumes increase, and the more operational risk is carried. Eventually, banks will re-engineer. What was innovative becomes boring, and is moved on to the core platform. By then, however, traders have innovated and there is a whole new set of innovative and exciting products, and the whole risk cycle continues.
Is it inevitable that something will go wrong, eventually?
In a way, yes. There is always innovation, and without innovation organisations die. If an institution sticks with spreadsheets, something will definitely go wrong, either because spreadsheets are so opaque and generally cause problems anyway, or through fraud. Criminals will realise that they can manipulate spreadsheets, that there’s no audit trail, and that nobody really knows how they work anyway—it’s easy just to change a few cells here and there.
Banks have to accept that innovation brings about innovation risk. They have to make sure they have the tools and processes in place that allow them to handle that innovation as quickly as possible, with a whole set of tools across the transaction lifecycle.
These controls can help firms to keep pace with innovation, so that if the front office innovates, the middle and back office can accommodate changes on the same day.
With CTC we are providing an enterprise solution to deal with volume and to be a single source of the truth. The solution should be completely audited and impossible to manipulate, with a view of the lifecycle and a certainty of the stage the transaction is at, making it as similar to the FedEx model as possible.
